www.gusucode.com > 粒子群工具箱 - Particle Swarm (PSO) Toolbox-beta-0.3源码程序 > 粒子群工具箱 - Particle Swarm (PSO) Toolbox-beta-0.3\get_psoOptions.m
%get_psoOptions >> A function to get an "options structure" that is used to set various option of the PSO Algorithm.
%
% Usage : psoOptions = get_psoOptions
% Arguments : None
% Return Values : psoOptions--> A Matlab structure. It is further divided into the following structures.
% |
% |_Flags------> PSO FLAGS. (All flags can be set to zero to disable and a positive value to enable)
% | |_ShowViz-----> Show visualization of the particles in motion. (suitable only if dimensions <= 3)
% | |_Neighbor---> Use neighborhood acceleration (in addition to global acceleration)
% |
% |_Vars------> PSO VARIABLES
% | |_SwarmSize--> Swarm Size. (Also known as population size)
% | |_Iterations-> Maximum Iterations. (Used to terminate the algorithm. see also: Terminate element below)
% | |_ErrGoal----> Error goal. (Also used to terminate the algorithm. see also: Terminate element below)
% | |_Dim--------> Dimensions of the problem. This determines the particle size.
% |
% |_SParams---> STRATERGY PARAMETERS
% | |_c1---------> Cognitive Acceleration
% | |_c2---------> Social Acceleration
% | |_c3---------> Neighborhood Acceleration
% | |_w_start----> Value of velocity Weight at the begining
% | |_w_end------> Value of velocity Weight at the end of the pso iterations
% | |_w_varyfor--> The fraction of maximum iterations, for which w is linearly varied
% | |_Vmax-------> Maximum velocity step
% | |_Nhood------> Neighborhood size (nhood=1 ==> 2 neighbors, one on each side)
% |
% |_Obj-------> OBJECTIVE FUNCTION OPTIONS
% | |_f2eval-----> Function/System to optimize (Type Fuzzy if u want to tune an FIS)
% | |_GM---------> Value of Global Minima (Required if Terminate.Err is set. i.e error goal is (one of) the termination criteria).
% | |_lb---------> Lower bounds of Initialization (You may use Asymmetrical Initialization. These bounds don't limit the search space.)
% | |_ub---------> Upper bounds of Initialization (Rather, they are used to Initialize the particles. See: An Empirical Study of PSO (Ebenhart and Shi)
% |
% |_Terminate-> TERMINATION OPTIONS (Multiple options are allowed)
% | |_Iters------> Use Vars.MaxIt as a termination criterion. 1=Yes, 0=No
% | |_Err--------> Use Vars.ErrGoal as a termination criterion. 1=Yes, 0=No
% |
% |_Disp------> TEXT MODE DISPLAY OPTIONS
% | |_Interval---> Notify about the progress after these many iterations. 0=Don't Notify
% | |_Header-----> Display the PSO Options header. 0 = Hide, 1 = Show
% | |_Progress---> Display the progress of the Algorithm. 0 = Hide, 1 = Show
% | |_Footer-----> Display Footer (contains experimental results). 0 = Hide, 1 = Show
% |
% |_Save------> OPTIONS FOR SAVING RESULTS IN A TEXT FILE
% | |_File-------------> Base name of the Output Text File. (see options below for automation options)
% | |_IncludeFnName----> Prefix the name of the name of the Objective function to the file name e.g. DeJong_Results. 1=Yes, 0=No
% | |_IncludeDim-------> Suffix the number of dimensions. DeJong_Results_10d. 1=Yes, 0=No
% | |_IncludeSwarmSize-> Suffix the size of swarm. DeJong_Results_10d20p. 1=Yes, 0=No
% | |_Interval---------> Save after this number of iterations. 0=Don't save the progress report.
% | |_Header-----------> Save the header information (also see: Disp)
% | |_Footer-----------> Save the Footer information (also see: Disp)
% |
% |_Fuzzy-----> FUZZY OPTIONS
% |_FIS-------------> An FIS file that contains the system to be optimize. (note: The actual file should have a .fis extension, although u may or may not include it here)
% |_DataFile--------> An valid Excel file containing Training Data. The FIS will be tuned to match this data.
% |_InitFunc--------> A function to initialize the Swarm using the FIS. (String or Function handle)
% |_ValidateFunc----> A function to Validate the Swarm each time particles are moved. This is required to change the co-ordinates of the particles which have bad value of membership function parameters.
% |_ErrorFunc-------> A function to evaluate the fitness of the fuzzy system. (note: Fitness is expressed in terms of mean square error. So, more fit functions would return a lower value of error)
% |_TuneInputs------> Tune the input membership functions
% |_TuneOutputs-----> Tune the output membership functions
%
% History : Author : JAG (Jagatpreet Singh)
% Created on : 06252003 (Wednesday. 25th June, 2003)
% Comments : Enjoy!
function psoOptions = get_psoOptions()
psoOptions = struct( ...
... %Option FLAGS
'Flags', struct(...
'ShowViz', 0, ... %Show visualization of the particles
'Neighbor', 0), ... %Use neighborhood acceleration (in addition to global acceleration)
... %Variables of the PSO
'Vars', struct(...
'SwarmSize', 20, ... %Swarm Size
'Iterations', 1000,... %Maximum Iterations
'ErrGoal', 1e-10, ... %Error goal
'Dim', 10 ), ... %Dimensions of the problem
... %Stratergy Parameters
'SParams', struct(...
'c1', 2, ... %Cognitive Acceleration
'c2', 2, ... %Social Acceleration
'c3', 1, ... %Neighborhood Acceleration
'w_start', 0.95, ... %Value of velocity Weight at the begining
'w_end', 0.4, ... %Value of velocity Weight at the end of the pso iterations
'w_varyfor',0.7,... %The fraction of maximum iterations, for which w is linearly varied
'Vmax', 100,... %Maximum velocity step
'Chi', 1, ... %Constriction factor
'Nhood', 1 ), ... %Neighborhood size (nhood=1 ==> 2 neighbors, one on each side)
... %Objective Function Options
'Obj', struct(...
'f2eval', 'DeJong', ...%Function/System to optimize
'GM', 0, ... %Value of Global Minima (Required if Error goal is a termination criteria).
'lb', 100, ... %Lower bounds of Initialization
'ub', 200 ), ... %Upper bounds of Initialization
... %Termination Options
'Terminate', struct(...
'Iters', 1, ... %Use Vars.MaxIt for termination
'Err', 1), ... %Use Vars.ErrGoal for termintion
... %Display Options
'Disp', struct( ...
'Interval', 10, ... %Notify about the progress after these many iterations
'Header', 1, ... %Display the PSO Options header. 0 = hide, 1 = show
'Progress', 1, ... %Display the progress of the Algorithm. 0 = hide, 1 = show
'Footer', 1 ), ... %Display Footer (contains experimental results). 0 = hide, 1 = show
... %Saving Options
'Save', struct(...
'File', 'Results', ... %Base name of the Output Text File. (see options below for automation options)
'IncludeFnName', 1, ... %Prefix the name of the name of the Objective function to the file name e.g. DeJong_Results
'IncludeDim', 1, ... %Suffix the number of dimensions. DeJong_Results_10d
'IncludeSwarmSize', 1, ... %Suffix the size of swarm. DeJong_Results_10d20p
'Interval', 10,... %Save after this number of iterations
'Header', 1, ... %Save the header information (see Disp)
'Footer', 1 ) , ... %Save the Footer information (see Disp)
... %info: the PSO Toolbox can used to tune the membership functions of an FIS. Set the options below.
... %Fuzzy Options
'Fuzzy', struct( ...
'FIS', 'tipper', ... %FIS file that contains the system to be optimize. (note: The file should end with a .fis extension)
'DataFile', 'tipperData', ... %An Excel File containing Training Data. The FIS will be tuned to match this data.
'InitFunc', @initfuzzy, ... %A function to initialize the Swarm using the FIS
'ValidateFunc', @validatefuzzy, ... %A function to Validate the Swarm each time particles are moved. This is required to change the co-ordinates of the particles which have bad value of membership function parameters.
'ErrorFunc', @fuzEvalFitness,... %A function to evaluate the fitness of the fuzzy system. (note: Fitness is expressed in terms of mean square error. So, more fit functions would return a lower value of error)
'TuneInputs', 1, ... %Tune the input membership functions
'TuneOutputs', 1) ... %Tune the output membership functions
);